INDEX

Explanations

negative values or concepts related to negativity

oai_token-act-pair · gpt-4o-mini Triggered by @bot

user input start

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Personendaten

-0.42

 nepri

-0.38

 counter

-0.36

enf

-0.36

vä

-0.35

เร

-0.34

 propOrder

-0.34

Bahnhof

-0.33

 waive

-0.33

uru

-0.33

POSITIVE LOGITS

tagHelperRunner

0.74

httphttps

0.66

 فريبيس

0.66

 resourceCulture

0.60

ValueStyle

0.56

parsedMessage

0.54

awtextra

0.54

 BorderSide

0.52

 ब्रेकडाउन

0.51

 للمعارف

0.49

Activations Density 0.000%

negative values or concepts related to negativity

user input start

No Comments

No Known Activations